"# LAPRAN-PyTorch"

This repository is an PyTorch implementation of the paper "LAPRAN: A Scalable Laplacian Pyramid Reconstructive Adversarial Network for Flexible Compressive Sensing Reconstruction", published in the Proceeding of the 15th European Conference on Computer Vision (ECCV) in 2018.

Code

Clone this repository into any place you want.

git clone
https://github.com/PSCLab-ASU/LAPRAN-PyTorch
cd LAPRAN-PyTorch

## Train your own model
You can start to train your own model via the following commands:

python main_adaptiveCS.py --model adaptiveCS_resnet_wy_ifusion_ufirst --dataset cifar10 --stage 1 --cr 20 --gpu 0

python main_adaptiveCS.py --model adaptiveCS_resnet_wy_ifusion_ufirst --dataset cifar10 --stage 2 --cr 20 --gpu 0

python main_adaptiveCS.py --model adaptiveCS_resnet_wy_ifusion_ufirst --dataset cifar10 --stage 3 --cr 20 --gpu 0

python main_adaptiveCS.py --model adaptiveCS_resnet_wy_ifusion_ufirst --dataset cifar10 --stage 4 --cr 20 --gpu 0

Then you get a four-stage LAPRAN for the cifar10 dataset.

## More codes of LAPRAN will be added to this repository later!

## Testing
The pretrained CIFAR10 model can be downloaded from: https://www.dropbox.com/s/eq2v2rowxqazj3u/results.zip?dl=0
Please download and upzip all files and directories to the root directory of the LAPRAN project.
. 
You can evaluate the pretrained model by running: python eval_adaptiveCS.py